Microstructure and hydrogen storage properties of the Mg2−xYxNi0.9Co0.1 (x = 0, 0.2, 0.3, and 0.4) alloys

Rare earth elements have excellent catalytic effects on improving hydrogen storage properties of the Mg2Ni-based alloys. This study used a small amount of Y to substitute Mg partially in Mg2Ni0.9Co0.1 and characterized and discussed the effects of Y on the solidification and de-/hydrogenation behaviors. The Mg2−xYxNi0.9Co0.1 (x = 0, 0.2, 0.3, and 0.4) hydrogen storage alloys were prepared using a metallurgy method. The phase composition of the alloys was studied using X-ray diffraction (XRD). Additionally, their microstructure and chemical composition were studied using scanning electron microscopy and energy-dispersive X-ray spectroscopy, respectively. The hydrogen absorption and desorption properties of the alloys were studied using pressure-composition isotherms and differential scanning calorimetric (DSC) measurements. The structure of the as-cast Mg2Ni0.9Co0.1 alloy was composed of the peritectic Mg2Ni, eutectic Mg–Mg2Ni, and a small amount of pre-precipitated Mg–Ni–Co ternary phases, and was converted into the Mg2NiH4, Mg2Ni0.9Co0.1H4, and MgH2 phases after hydrogen absorption. Furthermore, the XRD patterns of the alloys showed the MgYNi4 phase and a trace amount of the Y2O3 phase along with the Mg and Mg2Ni phases after the addition of Y. After hydrogen absorption, the phase of the alloys was composed of the Mg2NiH4, MgH2, MgYNi4, YH3, Y2O3, and Mg2NiH0.3 phases. With the increase of Y addition, the area ratios of the peritectic Mg2Ni matrix phase in the Mg2−xYxNi0.9Co0.1 (x = 0, 0.2, 0.3, and 0.4) alloys gradually decreased until they disappeared. However, the eutectic structure gradually increased, and the microstructures of the alloys were obviously refined. The addition of Y improves the activation performance of the alloys. The alloy only needed one cycle of de-/hydrogenation to complete the activation for x = 0.4. The DSC curves showed that the initial dehydrogenation temperatures of Mg2Ni0.9Co0.1 and Mg1.8Y0.2Ni0.9Co0.1 were 200 and 156 °C, respectively. The desorption activation energies of the hydrides of the Mg2Ni0.9Co0.1 and Mg1.8Y0.2Ni0.9Co0.1 alloys calculated using the Kissinger method were 94.7 and 56.5 kJ/mol, respectively. Moreover, the addition of Y reduced the initial desorption temperature of the alloys and improved their kinetic properties.

it to meet the requirements of the U.S. Department of Energy for hydrogen storage materials.Many methods showed significant improvement in the hydrogen storage performance of the Mg 2 Ni type alloys by substituting Ni or Mg with transition or rare earth elements, respectively, especially its thermodynamic properties could be adjusted to enable it to meet the requirements of practical applications 16 .Song et al. believed that partially substituting Mg with Nd could improve the activation property of the Mg 2 Ni alloy 17 .Consequently, it was reported for the first time that the hydrogen absorption capacity of Mg 1.9 Nd 0.1 Ni reached 2.86 wt%, which was higher than that of Mg 2 Ni since the multiphase structure formed by substituting Mg with Nd increased the phase boundary area and provided a favorable path for the diffusion of hydrogen atoms.Kalinichenka et al. showed that the Mg-Ni-Y system is highly suitable for reversible hydrogen storage 18 .The Mg 80 Ni 10 Y 10 and Mg 90 Ni 5 Y 5 alloys have a high hydrogen absorption rate under the hydrogen pressure of 20 bar, even at 250 °C.According to Li  et al. and Zhang et al., the addition of Y improved the hydrogen absorption and desorption thermodynamics of the MgNi-based alloys 19,20 .Xie et al. used the hydrogen plasma metal reaction (HPMR) method to successfully prepare Mg 2 Ni 1−x Co x (x = 0, 0.05, and 0.1) alloys and found that adding Co improved their hydrogen absorption kinetic properties significantly 21 .
Therefore, the literature review shows that the partial substitution of the A and B-side elements with Y and Co, respectively, contribute to improving the hydrogen storage performance of the Mg 2 Ni alloy [22][23][24][25] .However, few reports exist on the simultaneous addition of Y and Co.Thus, this study used the rare earth element Y and the transition element Co to replace Mg partially and Ni on the basis of Mg 2 Ni alloy, respectively, to realize the dual regulation of de-/hydrogenation kinetics and thermodynamic properties of the Mg 2 Ni alloy.

Experimental procedure
Preparation of the Mg 2−x Y x Ni 0.9 Co 0.1 samples Commercially pure Mg (99.9% purity), Mg-Ni intermediate alloy (70 wt% Ni content, 99.9% purity), Mg-Y intermediate alloy (30 wt% Y content, 99.9% purity), and pure Co (99.9% purity) were used as raw materials to prepare Mg 2−X Y x Ni 0.9 Co 0.1 (x = 0, 0.2, 0.3, and 0.4) alloy ingots in a graphite crucible in an electric resistance furnace, under the protection of a mixed flow of SF 6 and CO 2 .Intermittent mechanical agitation was conducted during smelting to prevent density segregation of the alloy melt.Moreover, additional Mg (2 wt%) was added to compensate for its inherent evaporation loss.Subsequently, the desired ingot could be obtained as the melt was cooled to room temperature in a furnace.The weight of each ingot was approximately 80 g.
Furthermore, the de-/hydrogenation measurement samples were taken from the center of each ingot.Prior to tests, the samples were mechanically broken, followed by ball-milling in a high-energy ball mill (HEBM) at a rotating speed of 240 rpm for about 60 min.The ball-to-material ratio during ball milling was 30:1, and highpurity argon (99.999% purity) was inserted to prevent oxidation.Additionally, intermittent ball milling was adopted to avoid the adhesion of components and excessively high temperature caused by long-duration ball milling.In intermittent ball milling, the rotation is stopped after 20 min for 15 min, then reversed until 60 min of ball milling is completed.Subsequently, about 0.5 g of 200-mesh powder was screened for the hydrogen storage performance test.

Characterization and de-/hydrogenation measurements
The phases in the as-cast ingots and hydrogenated powders were identified using X-ray diffraction (XRD) with Cu Kα radiation for continuous scanning at a rate of 2°/min in the 2θ range of 10-85°.The microstructures were characterized using scanning electron microscopy (SEM), and the corresponding chemical compositions were analyzed using energy-dispersive X-ray spectroscopy (EDS).Meanwhile, Image-Pro Plus (IPP) was used for counting the phase area ratios in the SEM images.
After activation, the isothermal de-/hydrogenation performance of the ball-milled alloys was measured using a precise volumetric Sieverts-type apparatus at 260, 280, and 300 °C, with hydrogen pressure of 2.5 and 0.1 MPa for hydrogen absorption and desorption, respectively.The differential scanning calorimetric (DSC) measurements were conducted from room temperature to 450 °C at the heating rates of 5, 10, and 15 °C/min under 50 mL/min argon gas flow to characterize the phase transformation behaviors of the hydrides.

Phase compositions of the as-cast alloys
The XRD patterns of the as-cast Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0, 0.2, 0.3, and 0.4) alloys are shown in Fig. 1.The diffraction peaks of the tested alloys were pointy, indicating their crystallization characteristics.Mg 2 Ni 0.9 Co 0.1 was only composed of the Mg and Mg 2 Ni phases.As shown in the Mg-Ni binary alloy diagram in Fig. 2, the solidification path of Mg 2 Ni under equilibrium solidification conditions should be: L → L + MgNi 2 → L + Mg 2 Ni → Mg + Mg 2 Ni.The absence of the MgNi 2 peaks in the Mg 2 Ni 0.9 Co 0.1 diffraction profile indicated that all the primary MgNi 2 was transformed into Mg 2 Ni though a peritectic reaction with liquid at 760 °C or the residual MgNi 2 was rare and diffused.Thus, the alloy was near-equilibrium solidified during furnace cooling in this study.
MgYNi 4 and Y 2 O 3 diffraction peaks were obviously present in the XRD patterns of Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0.2, 0.3, and 0.4) along with Mg and Mg 2 Ni diffraction peaks when Y was added to substitute Mg partially.MgYNi 4 belonged to Laves phase and had a C15 structure, which agreed well with Reference 26 .According to Reference 26 , the added Y was first dissolved in primary MgNi 2 to replace Mg, and then MgNi 2 reached the composition and transformed into MgYNi 4 when the Y content increased, and the Mg content decreased in MgNi 2 .Meanwhile, it can be inferred from the high diffraction intensity shown in Fig. 1 that the converted MgYNi 4 was largely absent from the subsequent peritectic reaction and was retained in the Mg 2−x Y x Ni 0.9 Co 0.1 alloy.

Microstructures of the as-cast alloys
The scanning electron microscope equipped with backscattered electrons detector (SEM/BSE) images of the as-cast Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0, 0.2, 0.3, and 0.4) alloys are shown in Fig. 3, and the corresponding EDS results are listed in Table 1.Based on the SEM images, EDS results, and the solidification path discussed, the as-cast Mg 2 Ni 0.9 Co 0.1 was composed of gray block peritectic Mg 2 Ni and dark lamellar eutectic Mg-Mg 2 Ni, with some bright fine primary MgNi(Co) 2 dispersed in the matrix, as shown in Fig. 3a and b.Meanwhile, the remaining  primary MgNi(Co) 2 was not found in the XRD test due to its small quantity and fine dispersion distribution characteristics.
As shown in Fig. 3c,e, and g, Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0.2, 0.3, and 0.4) alloys had largely increased amounts of the bright phase compared with Mg 2 Ni 0.9 Co 0.1 .Additionally, the corresponding EDS results showed that they were primary MgYNi 4 , coinciding well with the XRD results.Meanwhile, the original peritectic Mg 2 Ni block was refined and elongated after a small Y content (x = 0.2) was added, as shown in Fig. 3c.Furthermore, the primary precipitation phase was all MgYNi 4 , and the subsequent peritectic reaction was inhibited when the Y content increased above 0.In other words, the first precipitated MgNi 2 phase of the alloy was converted into MgYNi 4 under the condition of non-equilibrium solidification when the added Y content increased to 0.3, but MgYNi 4 did not participate in the peritectic reaction, leading to the precipitation of no peritectic Mg 2 Ni phase in the alloy.With the further decrease in temperature, the eutectic reaction occurred in the remaining alloy melt, forming the eutectic structure.Due to the high Y content in the melt, Y was dissolved in α-Mg in the eutectic structure to convert it to ε-Mg.Furthermore, the EDS analysis showed that the phase composition of the alloy did not change when the added Y content was increased to 0.4.Mg 1.6 Y 0.4 Ni 0.9 Co 0.1 and Mg 1.7 Y 0.3 Ni 0.9 Co 0.1 had the same solidification path, where MgYNi 4 was precipitated first, followed by the precipitation of Mg 2 Ni and ε-Mg through eutectic reactions.
The phase area ratios in the SEM images were counted using IPP to describe the effect of Y addition on the microstructure of the alloy more intuitively, and the corresponding statistically calculated results are shown in Fig. 4. Clearly, the peritectic reaction was inhibited when the added Y content was above 0.2.Compared with Mg 2 Ni 0.9 Co 0.1 , the area ratio of the peritectic Mg 2 Ni decreased from 67 to 21% when a small amount of Y (x = 0.2) was added.Meanwhile, more melt was retained at low temperatures to conduct the eutectic reaction since the peritectic reaction was inhibited, increasing the area ratio of eutectic Mg-Mg 2 Ni with increased Y content.The maximum area ratio for MgYNi 4 was obtained when the added Y content was 0.3.Thus, the added Y was first dissolved in primary MgNi 2 , which was converted into MgYNi 4 when Y reached the corresponding content in the Mg 2−x Y x Ni 0.9 Co 0.1 alloy.The solidification path was L  converted into MgYNi4 when the added Y content was above 0.2, inhibiting the peritectic reaction and disappearance of peritectic Mg 2 Ni in the final solidified structures.Thus, the MgYNi 4 amount in the final solidified microstructure depended on the amounts of primary Mg(Y)Ni2 and Y that could participate in the reaction during solidification.The amount of the primary Mg(Y)Ni 2 decreased when the added Y content was further increased from 0.3 to 0.4, reducing the final transformed MgYNi 4 , as shown in Fig. 4. The excess Y will dissolved into Mg-Mg 2 Ni eutectic during subsequent solidification, resulting in the higher Y content dissolved in eutectic Mg-Mg 2 Ni in Mg 1.6 Y 0.4 Ni 0.9 Co 0.1 than that in Mg 1.7 Y 0.3 Ni 0.9 Co 0.1 and Mg 1.8 Y 0.2 Ni 0.9 Co 0.1 , as shown in Table 1.In addition to the phase composition discussed, adding Y significantly refined the solidification structure, as seen in Fig. 3.

Activation and de-/hydrogenation properties
Since an impurity layer composed of oxides and hydroxides is usually formed on the surface of ball-milled alloy particles, activation is necessary to break this surface impurity layer and expose the fresh alloy metal inside.This study carried out the activation through three successive de-/hydrogenation circles at 300 °C.Hydrogenation and dehydrogenation were conducted under 2. www.nature.com/scientificreports/temperature of YH 2 is about 1063 K 27 , which is obviously much higher than the activation temperature in this study.Consequently, YH 2 could not be dehydrogenated and was left, resulting in the decreased hydrogen absorption capacity of the alloy in the subsequent hydrogen absorption process.Meanwhile, Mg 1.6 Y 0.4 Ni 0.9 Co 0.1 had a more obvious reduction in the hydrogen absorption capacity after activation since Mg 1.6 Y 0.4 Ni 0.9 Co 0.1 had a larger Y content than Mg 1.8 Y 0.2 Ni 0.9 Co 0.1 , as shown in Fig. 5b and c.Due to its low average atomic density and the large space between atoms, hydrogen diffused more easily through the phase boundary 28 .As discussed in Section "Microstructures of the as-cast alloys", the area ratio of eutectic structures increased, and the solidification structure was refined with the increased Y content, improving the activation property of the alloys.
The hydrogen absorption capacity and rate at different temperatures are important indexes to reflect the hydrogen absorption kinetics of the alloys.The hydrogen absorption kinetic curves of Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0, 0.2, and 0.4) alloys at different temperatures are shown in Fig. 6.All the three alloys showed a fast hydrogen absorption rate after being fully activated, reaching more than 80% of the corresponding maximum hydrogen absorption capacity in only 6 min.As shown in Section "Microstructures of the as-cast alloys", MgYNi 4 was formed when Y was added, and its proportion increased with the increased Y content.The maximum hydrogen absorption capacity of Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0, 0.2, and 0.4) decreased with the increased Y content, from 3.31 wt% to 1.99 wt% and then to 1.67 wt% at 260 °C due to the presence of the non-hydrogen absorbing phase.
The maximum hydrogen absorption capacity of the Mg 2 Ni 0.9 Co 0.1 alloy was 3.13 wt% at 2.5 MPa hydrogen pressure and 300 °C temperature, and the time taken to reach the maximum hydrogen absorption capacity was 22 min.Under the same reaction conditions, Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0.2 and 0.4) alloys could reach more than 90% of the maximum hydrogen absorption capacity in 200 s.By comparison, the maximum hydrogen absorption of the alloys decreased with the addition of Y due to the formation of the MgYNi 4 phase that does not absorb hydrogen and the unsaturated hydride Mg 2 NiH 0.3 in the Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0.2 and 0.4) alloys during the hydrogen absorption process.
The hydrogen desorption kinetic curves of Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0, 0.2, and 0.4) alloys at 0.1 MPa hydrogen pressure after hydrogen absorption under the conditions mentioned are shown in Fig. 7.The comparison of the de-/hydrogenation kinetic curves of the alloys showed that the complete hydrogen desorption of alloys required significantly lower time than the hydrogen absorption process.The hydride layer on the surface of the alloys hinders the diffusion of hydrogen atoms into the alloys during their hydrogen absorption process.However, this hydride layer on the surface of the alloys breaks down during the desorption process to produce hydrogen atoms that do not need to pass through the hydride layer but only through the metal surface.Since the diffusion rate 0 600 1200 1800 2400 3000 3600 of the hydrogen atoms in the hydride was much lower than that in the metal surface 29 , the complete hydrogen desorption of the alloys required significantly lower time than the complete hydrogen absorption.Additionally, the maximum hydrogen discharges of the three alloys increased with the increase in temperature since the dehydrogenation of the alloy is a reversible process and the hydrogen absorption and desorption reactions occur simultaneously, where the absorption and desorption processes are exothermic and endothermic reactions, respectively.The hydrogen desorption of the alloys would increase with the increase in temperature since the temperature is conducive to the desorption process.The dehydrogenation capacities of the Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0, 0.2, and 0.4) alloys at different temperatures for 50 s are shown in Table 2.The dehydrogenation amounts of the alloys for x = 0, 0.2, and 0.4 were 0.19, 0.22, and 0.24 wt% after hydrogen desorption for 50 s at 260 °C, respectively, indicating that the addition of Y improved the desorption kinetic properties of the alloys.However, this phenomenon was not observed at higher reaction temperatures.

Hydrogen absorption reaction mechanism of alloys
The XRD patterns of the Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0, 0.2, and 0.4) alloy hydrides after hydrogen absorption at 300 °C and 2.  Additionally, no other Co phases except the Mg 2 Ni 0.9 Co 0.1 H 4 phase were found in the XRD pattern of the Mg 2 Ni 0.9 Co 0.1 alloy hydride.Therefore, it could be concluded that the Mg-Ni-Co phase did not undergo hydrogen absorption or decomposition reaction under the experimental conditions of this study, which is consistent with the fact that MgNi 2 in the Mg 2 Ni alloy did not undergo the hydrogen absorption reaction 31 .
The XRD patterns of the Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0.2 and 0.4) alloys after hydrogen absorption showed that both of them were composed of Since the samples were ball-milled in the argon atmosphere and the operation of taking out the samples and loading them into the reactor was performed in a glove box, the trace Y element in the alloy would not be completely oxidized, and YH 3 would be formed in the hydrogen absorption process.The intensity of the YH 3 diffraction peak increased slightly with the increased Y content, indicating that more Y was involved in the reaction to form YH 3 in the hydrogen absorption process of the alloys with the increased substitution of Y.
Combined with the XRD patterns and EDS analysis, the hydrogen absorption reaction of the Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0.2 and 0.4) alloys can be expressed as follows: (1) The DSC curves of the Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0 and 0.2) alloy hydrides at different heating rates of 5, 10, and 15 °C/min are shown in Fig. 10.The peak desorption temperatures of the alloys at different heating rates are presented in Table 3.The heating rate affected the peak desorption temperature of the alloy hydrides.The peak desorption temperature increased for both alloys with the increase in the heating rate.The dehydrogenation activation energy E a is the height of the barrier between the lowest potential energy of the dehydrogenation reactant and the product.The kinetic properties of hydrogen desorption were better for lower activation energies.Furthermore, this study used the peak temperatures of the DSC curves with different heating rates to investigate the dehydrogenation kinetics of the Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0 and 0.2) alloys.The dehydrogenation activation energy of the Mg 2−x Y x Ni 0.9 Co 0.1 (x = 0 and 0.2) alloy was calculated according to the Kissinger equation shown in Eq. 8.
where, β , T P , E a , R, and A represent the linear heating rate (°C/min or K/min), the dehydrogenation peak tem- perature (K), the activation energy for hydrogen desorption (kJ/mol), the gas constant (8.314J/mol/K), and a linear constant, respectively.
As shown in Fig. 11, the slope of the plot of ln β/T 2 P versus 1/T P was E a /R.Using the slope of the straight line, the calculated dehydrogenation activation energy of the Mg 2 Ni 0.9 Co 0.1 alloy was 94.7 kJ/mol, slightly lower
3, resulting in the disappearance of peritectic Mg 2 Ni.Consequently, the Mg 1.7 Y 0.3 Ni 0.9 Co 0.1 and Mg 1.6 Y 0.4 Ni 0.9 Co 0.1 alloys were only composed of bright primary MgYNi 4 and dark lamellar eutectic Mg-Mg 2 Ni.Unlike the Mg 2 Ni 0.9 Co 0.1 alloy, the Mg 1.8 Y 0.2 Ni 0.9 Co 0.1 alloy had four regions with contrasting degrees.The results obtained using EDS analysis of the composition of each region are shown in Table 1.Combined with the XRD results of the alloys, it could be seen that the bright white region A embedded in the matrix was the first precipitation of the MgYNi 4 phase.The elongated acicular gray region B was the eutectic Mg 2 Ni phase, the black region C around the acicular Mg 2 Ni phase was the eutectic α-Mg phase (Mg phase without a solid solution of other elements), and the gray-black region D was the eutectic ε-Mg phase (Mg phase with small amounts of Ni

Figure 3 .
Figure 3. SEM/BSE images of the as-cast Mg 2−x Y x Ni 0.9 Co 0.1 alloys at low and high magnification for (a and b) x = 0, (c and d) x = 0.2, (e and f) x = 0.3, and (g and h) x = 0.4.